Greece since the 1960s: the mortality transition revisited: a joinpoint regression analysis

Abstract

Mortality transition in Greece is a well-studied phenomenon in several of its aspects. It is characterised by an almost constant increase in life expectancy at birth and other ages and a parallel decrease in death probabilities. The scope of this paper is a comprehensive assessment of the mortality transition in Greece since 1961, in the light of holistic analysis. Within this paper, life tables by gender were calculated and the temporal trends of life expectancy at several ages were examined. Moreover, a cluster analysis was used in order to verify the temporal changes in the mortality patterns. The probabilities of death in large age classes are presented. Furthermore, the death distribution was analysed in relation to various parameters: the modal age at death, mode, left and right inflexion points and the length of the old age heap. Before that, a non-linear regression method, originating from the stochastic analysis, was applied. Additionally, the Gini coefficient, average inter-individual differences, and interquartile range of survival curves were examined. Finally, the standardised rates of the major causes of death are presented. All the analysis variables were scholastically examined for their temporal trends with the method of Joinpoint Regression analysis. Mortality transition in Greece after the year 1961 is asymmetrical with a gender and an age-specific component, leading to the elevation of life expectancy at birth over time. During this period, the older ages' mortality decreases, but at a slower pace than that of the younger ones. The modal age at death, mode, the left and right inflexion points and the width of the old age heap denote the compression of mortality in the country. The old age death heap shifts towards older ages, while at the same time, the variability of ages at death decreases, being verified by the Gini Coefficient and average inter-individual differences. As a result, the rectangularization of survival curves is evident. These changes have a different pace of transition over time, especially after the emergence of the economic crisis. Finally, the major causes of death were the diseases of the circulatory system, neoplasms, diseases of the respiratory system and others. The temporal trends of these diseases differ according to the diseases and gender. Greece's mortality transition is an asymmetrical stepwise process characterised by its gender and age-specific characteristics. This process, despite being a continuous one, is not linear. Instead, a combination of serious developments over time governs the country's modern mortality regime. The evaluation of Greece's mortality transition through the lens of more advanced analytical methods may provide new insights and methodological alternatives for assessing mortality transition in other countries of the world.

Keywords: Cluster analysis; Gini coefficient; Greece; Joinpoint analysis; Left and right inflexion points; Life expectancy; Modal age at death; Mode; Mortality transition; Non-linear regression; Probabilities of death.

Fig. 1

Left and right inflexion points. Source: Zafeiris and Skiadas (2017)

Fig. 2

Life expectancy at birth, 15 and 30 years. Greece 1961–2019. AAPC values are stored in row “Overall” and column “APC”

Fig. 3

Decomposition of life expectancy at birth (e0) in large age classes. Averages for time periods

Fig. 4

Life expectancy at ages 45 and 65 years. Greece 1961–2019. AAPC values are stored in row “Overall” and column “APC”

Fig. 5

Cluster analysis of the mortality patterns. Cophenetic correlation coefficient: Males = 0.7 Females = 0.8

Fig. 6

Probabilities of death at ages < 30. *Asterisk denotes statistical significance at 0.05 level

Fig. 7

Probabilities of death at ages >  = 30. *Asterisk denotes statistical significance at 0.05 level

Fig. 8

The death density probability distribution. Greece 1986–2019

Fig. 9

Modal age at death and Mode of the death density probability distribution. Greece 1986–2019. AAPC values are stored in row “Overall” and column “APC”

Fig. 10

Left and right inflexion points and width of the death distribution in the old age death heap. Greece 1986–2019. AAPC values are stored in row “Overall” and column “APC”

Fig. 11

The Gini coefficient (index) and the average inter-individual difference (AID). Greece 1986–2019. AAPC values are stored in row “Overall” and column “APC”

Fig. 12

The life expectancy at birth (e0) and Gini coefficient (index). Greece 1986–2019

Fig. 13

The survival curves (lx). Greece 1986–2019

Fig. 14

The quartiles of the survival curve and the interquartile range. Greece 1986–2019. AAPC values are stored in row “Overall” and column “APC.”

Fig. 15

The major causes of death. Directly standardised rates. Greece 1994–2019